Missing TNM concepts

[gkennos]

Missing term	Parent	Valid because
p-8th_aJCC/UICC-N0(i-)	https://athena.ohdsi.org/search-terms/terms/1635560	pN0(i-) exists in NAACCR and also in the AJCC documentation (also in the data)
p-8th_aJCC/UICC-N0(mol-)	https://athena.ohdsi.org/search-terms/terms/1635560	Present in data as well as SEER docs and AJCC docs, though SNOMED term has been invalidated so validity TBC
p-8th_aJCC/UICC-Tis(LCIS)	https://athena.ohdsi.org/search-terms/terms/1633920	As per (mol-)

Also true for other editions. (i-) is the most prevalent and most supported via other terms, but the others also present in standards and in the data, so direction required.

[cgreich]

The (i-), (i+), (mol-) and (mol+) make sense as children of N0. The DCIS/LCIS right now is part of ICDO3 (e.g. this and this). We probably don't want it twice.

[gkennos]

DCIS is already in the cancer modifier vocab, as well as being in the staging standard so it feels a bit inconsistent to leave LCIS out?

1635835

[rtmill]

@kzollove appropriately added this to the TNM task group above.

The (i-), (i+), (mol-) and (mol+) make sense as children of N0. The DCIS/LCIS right now is part of ICDO3 (e.g. this and this). We probably don't want it twice.

@cgreich I'm not seeing the overlap and consequently concern here. We're specifically referring to staging observations without the context of the related condition. Can you elaborate on why the existence of ICDO3 codes is needing to be considered here?

[cgreich]

@rtmill: Sure.

The problem is how to handle the "in situ" cancers. On one hand, they are early stage. On the other hand, because they are early stage, they are not visible macroscopically and therefore reported as histological findings. So, is the "in situ" a stage (and hence a Cancer Modifier Measurement) or a histology (and hence a ICDO3 Condition)? Can't have it both ways in our system.

I know the oncologists use both ways. They don't have a precise model to work with that has to adhere to the principle of one representation of fact.

There are two solutions:

• We add the AJCC concepts containing LCIS (and DCIS also, btw), but make them non-standard and map them over.
• We do the opposite, de-standardize the ICDO3s and map them over to the AJCC concepts.

The former is probably easier than the latter. But then the hierarchy won't pick them up.

• Or we have both.

Which means we kick the can down to the analyst who has to realize the duplication and do the queries appropriately. It sounds worse than it is in my opinion, because the analyst already has to realize that "metastatic disease" means looking for Metastasis and all its descendants as well as for AJCC/UICC M1 Category and all its descendants.

So, maybe yes, let's add it.